Frequently asked questions
Quick answers about Mithrify and early access.
What is Mithrify?
Mithrify is the local, user-owned memory layer for AI work. It preserves AI conversations, prompts, decisions, project reasoning, code-agent sessions, and useful outputs; keeps the original source locally; crystallizes durable knowledge into Markdown; and makes that memory searchable and reusable.
Is Mithrify another chatbot?
No. Mithrify is not trying to replace the AI tools you already use. Its primary role is preserving, structuring, and recalling the work produced across those tools so useful context does not disappear between sessions.
Is this a second brain or note-taking app?
Not in the traditional sense. You can create and edit Markdown in Mithrify, but the product is designed to reduce manual documentation by turning existing AI conversations and agent sessions into durable working memory. Notes are editable knowledge and a useful output of crystallization, not a constant documentation chore.
Which AI tools does it support?
Claude Code is the first importer priority. Import support for ChatGPT, Claude, and Cursor is planned to roll out progressively during early access. The underlying memory model is provider-independent so Mithrify does not depend on any one AI vendor.
Is everything local?
Mithrify is local-first by default. Canonical knowledge is stored as readable Markdown, while original conversations, session records, attachments, provenance, indexes, and operational metadata remain locally controlled. Mithrify does not require a cloud service to own your memory.
What is crystallization?
Crystallization turns a long AI conversation or agent session into concise, durable Markdown containing the decisions, reasoning, solutions, prompts, tasks, unresolved questions, and other useful outcomes. The original source remains preserved so the crystallized result can be traced back to its evidence.
How does Mithrify organize ongoing AI work?
Mithrify preserves complete Conversations, divides work into meaningful Sessions, and connects related work through Topics. This lets project continuity survive across separate chats, tools, and working periods instead of reducing everything to isolated notes.
Will AI agents be able to retrieve Mithrify memory?
That is the direction. Mithrify's planned read-only MCP Memory Server will allow compatible agents such as Claude Code and Cursor to search and retrieve approved local memory. MCP access is a roadmap capability and will follow reliable archive, crystallization, and search foundations.
Does Mithrify replace Claude, Cursor, or ChatGPT?
No. You keep using the AI tools that work for you. Mithrify provides a provider-independent memory layer beneath them so your working context is not trapped inside one vendor, product, or session.
Can I leave Mithrify without losing my work?
Yes. Durable textual knowledge remains readable Markdown, with ordinary attachments and exportable source records. Mithrify is designed so your knowledge remains useful without requiring permanent access to Mithrify or a specific AI provider.
Is this for teams or individuals?
Early access focuses on individuals: founders, developers, operators, and researchers who work with AI regularly. Collaborative and team workflows are a later direction.
What is early access?
Mithrify is being developed toward an early-access desktop release. Join the list for an opportunity to access early builds and help determine which importers and workflows receive priority.
Join Early Access
Import your AI work. Crystallize what matters. Recall it when the next session starts.